Pāriet uz galveno navigāciju Pāriet uz meklēšanu Pāriet uz galveno saturu

Creating Lithuanian and Latvian Speech Corpora from Inaccurately Annotated Web Data

  • Tilde Company

Zinātniskās darbības rezultāts: Nodaļa grāmatā/enciklopēdijā/konferences krājumāKonferences zinātniskais rakstsPētniecībakoleģiāli recenzēts

8 Atsauces (Scopus)

Kopsavilkums

This paper describes the method that was used to produce additional acoustic model training data for the less-resourced languages of Lithuanian and Latvian. The method uses existing baseline speech recognition systems for Latvian and Lithuanian to align audio data from the Web with imprecise non-normalised transcripts. From 690 hours of Web data (300h for Latvian, 390h for Lithuanian), we have created additional 378 hours of training data (186h for Latvian and 192 for Lithuanian). Combining this additional data with baseline training data allowed to significantly improve word error rate for Lithuanian from 40% to 23%. Word error rate for the Latvian system was improved from 19% to 17%.

OriģinālvalodaAngļu
Rīkotāja publikācijas nosaukums11th International Conference on Language Resources and Evaluation (LREC 2018), May 7-12, 2018, Miyazaki, Japan : Conference Paper
RedaktoriNicoletta Calzolari, Khalid Choukri, Christopher Cieri, Thierry Declerck, Sara Goggi, Koiti Hasida, Hitoshi Isahara, Bente Maegaard, Joseph Mariani, Helene Mazo, Asuncion Moreno, Jan Odijk, Stelios Piperidis, Takenobu Tokunaga
IzdevējsEuropean Language Resources Association (ELRA)
Lapas2871-2875
ISBN (Drukātā versija)979-109554600-9
Publikācijas statussPublicēts - 2018
Pasākums11th International Conference on Language Resources and Evaluation, LREC 2018 - Miyazaki, Japāna
Ilgums: 7 maijs 201812 maijs 2018

Publikāciju sērijas

NosaukumsLREC 2018 - 11th International Conference on Language Resources and Evaluation

Konference

Konference11th International Conference on Language Resources and Evaluation, LREC 2018
Valsts/TeritorijaJapāna
PilsētaMiyazaki
Periods7/05/1812/05/18

OECD Zinātnes nozare

  • 1.2 Datorzinātne un informātika

Nospiedums

Uzziniet vairāk par pētniecības tēmām “Creating Lithuanian and Latvian Speech Corpora from Inaccurately Annotated Web Data”. Kopā tie veido unikālu nospiedumu.

Citēt šo